Port driven authentication in a network

Abstract

A method for authenticating a message in a time division multiple access network is provided. The method includes receiving a message from an active relaying component, inspecting a value in the message inserted by the active relaying component, and comparing the value with an expected value based on a transmission schedule.

Description

CROSS REFERENCE TO RELATED APPLICATIONS

This application is related to and claims the benefit of the filing date of the following U.S. Provisional Applications:

Ser. No. 60/560,323, filed on Apr. 6, 2004 and entitled “MESSAGE AUTHENTICATION IN A COMMUNICATION NETWORK” (the '323 Application).

Ser. No. 60/523,785, filed on Nov. 19, 2003 and entitled “PRIORITY BASED ARBITRATION FOR TDMA SCHEDULE ENFORCEMENT IN A DUAL LINK SYSTEM” (the '785 Application).

Ser. No. 60/523,782, filed on Nov. 19, 2003 and entitled “HUB WITH INDEPENDENT TIME SYNCHRONIZATION” (the 782′Application).

These provisional applications are incorporated herein by reference.

This application is also related to the following US non-provisional applications:

Attorney docket no. H0005459 (400.010US01) entitled “PRIORITY BASED ARBITRATION FOR TDMA SCHEDULE ENFORCEMENT IN A MULTI-CHANNEL SYSTEM” filed on even date herewith (the '549 Application).

Attorney docket no. H0005031 (400.005US01) entitled “ASYNCHRONOUS HUB” filed on even date herewith (the '031 Application).

These non-provisional applications are incorporated herein by reference.

BACKGROUND INFORMATION

Communication networks are used in a variety of applications including telephone and computer systems, weapons systems, navigational systems, and advanced control systems in cars, aircraft and other complex systems. Given the variety of applications, many kinds of communications networks have been developed over the years. One common characteristic of communication networks is the use of a communication medium that interconnects various nodes on the network. Various topologies and protocols have been developed to control communications between the nodes of these networks.

One network architecture is referred to as Time Division Multiple Access (TDMA). In a TDMA network, nodes in the network are assigned time slots for communicating over the network. Many different TDMA protocols have been developed for communication between nodes of a network. For example, these protocols include TTP/C, SAFEbus, FlexRay and other TDMA protocols.

FIG. 1 is a block diagram of a conventional communication network 100. Network 100 includes a plurality of communicating end stations, referred to herein as nodes, (1). The network 100 also includes relaying devices (2), commonly referred to as guardians. Each guardian (2) has a number of ports (5). In this example, network 100 includes two guardians (2). This allows all data from one node (1) to be transmitted to another node over two different channels, e.g., one channel per guardian. Two channels are commonly used in the industry to provide fault tolerance for the network 100. In other embodiments, any appropriate number of guardians (2) may be used.

The network 100 also includes a number of communication links (3). Communication links (3) comprise one or more of wired, e.g., copper cable, twisted pair, coaxial cable, optical fiber; wireless, e.g., radio frequency, infrared; or other appropriate communication medium. The communication links (3) couple each node (1) with each guardian (2). Messages (4) are sent between nodes (1) through the guardians (2).

Collectively, the nodes (1), guardians (2), and communication links (3) comprise a cluster.

One problem with these TDMA based networks is that faulty nodes can masquerade as other nodes. Masquerading nodes can influence local decisions by pretending to be another node and by behaving differently on different communication channels.

Fault-tolerant systems with only two replicated communication channels are vulnerable to masquerading-induced failure. Previous techniques developed to mitigate masquerade failures require additional redundancy (adding more nodes) and global knowledge of the communication schedule embodied and enforced by a central guardian. In some conventional systems, these techniques have been achieved using dedicated special communication links between guardians. Unfortunately, this, in turn, has compromised the underlying requirement of channel independence in the fault tolerant network. Furthermore, the existing techniques developed to date assume that benign failure modes (such as “stuck at” faults in a guardian). There are other simple failure modes that can cause total system failure, such as shorts between a guardian's inputs and outputs that destroys a network's clock synchronization mechanism by eliminating the assumed propagation delay through each guardian. Further, before achieving synchronous operation, authentication assumed by TDMA protocols cannot be counted on due to the possibility of the masquerading-induced failures and hence proper start-up is also an issue.

Therefore, a need exists for an improved technique for containing faults caused by masquerading nodes in a time division multiple access network.

SUMMARY

Embodiments of the present invention overcome problems with masquerading nodes by requiring the guardian to take an affirmative action when passing data between nodes. In one embodiment, the guardian writes a value, e.g., derived from the port number associated with the node sending the data, into a field in the data. In other embodiments, the guardian performs other actions such as inverting the data from the node, pre-pending or post-pending a value derived from the port number, or other appropriate action that allows the receiving node to authenticate the origin of the data.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of a conventional star architecture with dual channels and two guardians.

FIG. 2 is a data flow diagram that illustrates one embodiment of the flow of data in a network using independent port-derived authentication.

FIG. 3 is a data flow diagram that illustrates another embodiment of the flow of data in a network using independent port-derived authentication.

DETAILED DESCRIPTION

Embodiments of the present invention address problems with masquerading nodes and with faulty guardians having by-passes or short circuits. Embodiments of the present invention overcome these problems by requiring the guardian to perform an active behavior when relaying traffic.

FIG. 2 is a data flow diagram that illustrates the flow of data in a network 200 using independent port-derived authentication. In this example, network 200 has four nodes identified as Node 1, Node 2, Node 3, and Node 4. Network 200 also includes an active relaying component 202. It is understood that the number of nodes and active relaying components are shown by way of example and not by way of limitation. The techniques described herein are appropriately applied in networks with any appropriate number of nodes and active relaying components. Each node is coupled to active relaying component 202 over a communication medium 204, e.g., one or more of wired, e.g., copper cable, twisted pair, coaxial cable, optical fiber; wireless, e.g., radio frequency, infrared; or other appropriate communication medium.

In operation, network 200 uses port-driven authentication to contain faults in the network. In one embodiment, network 200 uses independent port-derived authentication (IPDA) to contain faults. In one embodiment, the active relaying component 202 writes the port number from which it is relaying the message into the relayed message data stream. In other embodiments, the active relaying component 202 writes a value based on the port number to be written into the relayed message data stream. In other embodiments, other network-wide, unique identifiers are used that identify the source in the network of the message. This active behavior allows detection of fail-passive behavior of the hub and as well as the detection of masquerading nodes by authorizing a message source by an independent device (the active relaying component 202). As this action is based on ports, the active relaying component 202 does not require schedule knowledge.

Advantageously, knowledge of an authorized message can be leveraged in several protocol algorithms deployed in the nodes and can simplify algorithm complexity and replication requirements, while allowing fully independent communication channels and active relaying components, e.g., guardians, and without the active relaying components having knowledge on the TDMA schedule.

To implement this technique, the active relaying component 202 takes a known action on data from the nodes. For example, in one embodiment, the message from a node includes a special field, called the port-tag field. Each node in network 200 is assigned a unique number in network 200. In one embodiment, the unique number is based on the port of the active relaying component 202 to which the node is attached. Further, each node in the network 200 has knowledge of the architectural mapping (that is, which node is attached to which port on active relaying component 202).

The independent port-derived authentication (IPDA) technique implemented in network 200 identifies a relayed frame (assuming a correct sender) as incorrect (by each receiving node) when the action of the active relaying component 202 is missing, e.g., by-passed by a short-circuit between the input and outputs of the active relaying component 202 either internal or external to the active relaying component 202. The IPDA technique assumes that the active relaying component 202 will not perform the required correct action if the active relaying component 202 is faulty. Furthermore, each node authenticates received messages based on knowledge of the other nodes' port numbers and the transmission schedule of the nodes in network 200. Thus, in one embodiment, each node evaluates the port-tag of a received data or message in combination with knowledge of the transmission schedule to verify that the data came from the correct node.

In operation, network 200 implements the IDPA technique to authenticate data relayed by active relaying component 202. Each node includes a reserved field, e.g., the port tag field, within every transmission it makes. In one embodiment, the nodes set the value of this field to a default port tag value (one that is different from any port number present in the system) when transmitting data to the active relaying component 202. For example, as shown in FIG. 2, the port tag field in the message 206 is set to a value of 111. This value is chosen as the default value and is not associated with any port of active relaying component 202.

Active relaying component 202 over-writes the port tag in the frame as the transmission is relayed through it. In one embodiment, the value placed in the reserved field is based on the port from which the active relaying component 202 receives the frame. Continuing with the example in FIG. 2, when Node 1 sends a frame, the active relaying component 202 forwards the frame to all other nodes after inserting the port tag of 101 for port 5 into the frame in the port tag field as indicated at 208 thus indicating that the message originated at Node 1.

Each node receives and authenticates the message from the active relaying component 202. When the frame is received, the node inspects the port tag in port tag field 208. The node compares the received port tag with an expected port tag derived from a known transmission schedule in table 210. Using this information, the node can determine if the frame stems from a correct or a masquerading node, and if the active relaying component is properly performing its tagging function. Again, continuing the example in FIG. 2, node 3 receives the frame from active relaying component 202. Node 3 compares the port tag value of 101 with the value in table 210 stored in Node 3 to determine if the message originated from the correct node and if the active relaying component is functioning properly. Since all information at a receiving node is based an a priori knowledge, active relaying component 202 authorizes the node of origin of a message.

Advantageously, when the active relaying component 202 uses the port number in the active relaying component as a network-wide unique identifier, the complexity of the active relaying component 202 can be minimized as the active relaying component 202 does not require schedule knowledge, if suitable arbitration schemes are deployed as described, for example, in the '785 or '549 Applications.

FIG. 3 is a data flow diagram that illustrates another embodiment of the flow of data in a network 300 using independent port-derived authentication. In this embodiment, the node that originates the message puts the correct port number into the message. The active relaying component 302 checks that the port number in each message matches the port that the message arrived on. This has the benefit that the port number can be included in a message's error detection mechanism (e.g. CRC or checksum). The active relaying component 302 must still do something to the message to allow receiving nodes to detect faults that by-pass active relaying component's checking.

In this embodiment, node 1 through node N each includes a respective one of inverters 320-1 to 320-N. The inverters 320-1 to 320-N invert the message for either transmission or reception, but not both. The active relaying component 302 also includes an inverter 322. The active relaying component also inverts the relayed messages. With no fault, each message is inverted twice, which leaves the message in its initial form. If the active relaying component 302 is by-passed, the message is only singly inverted, which would cause a node's receiver to flag such messages as erroneous. In addition to inverting the message, in one embodiment, the active relaying component 302 also takes some action based on local information to provide the message authentication function. In one embodiment, this action uses the port number as the local information.

In another embodiment of network 200 of FIG. 2, the active relaying component appends a port-derived value, e.g., a port tag, to one end of each relayed message. In one embodiment, the active relaying component pre-pends the value to the leading edge of the message. In another embodiment, the active relaying component post-pends the value to a trailing edge of the message. This has the benefit that any pre-existing transmission mechanism(s) would not have to be changed to add the IPDA technique.

The IPDA technique also could be used in multi-hop networks in which a message is relayed sequentially through 2 or more active relaying components. One method to do that would appear to be the inverse of wormhole routing—each hop through an active relaying component would add another port-tag to the message.

Advantageously, the IPDA technique provides a low-cost communication schemes for networks that need (or assume) authentication. Further, the IPDA technique also may greatly simplify start-up in TDMA schemes when traditional TDMA-based authentication cannot be assumed. The deployment of this IPDA technique may also greatly simplify the design of guardian implementations, as it removes the need of schedule knowledge and associated programming logistics.

The methods and techniques described here may be implemented in digital electronic circuitry, or with a programmable processor (for example, a special-purpose processor or a general-purpose processor such as a computer) firmware, software, or in combinations of them. Apparatus embodying these techniques may include appropriate input and output devices, a programmable processor, and a storage medium tangibly embodying program instructions for execution by the programmable processor. A process embodying these techniques may be performed by a programmable processor executing a program of instructions stored on a machine readable medium to perform desired functions by operating on input data and generating appropriate output. The techniques may advantageously be implemented in one or more programs that are executable on a programmable system including at least one programmable processor coupled to receive data and instructions from, and to transmit data and instructions to, a data storage system, at least one input device, and at least one output device. Generally, a processor will receive instructions and data from a read-only memory and/or a random access memory. Storage devices or machine readable medium suitable for tangibly embodying computer program instructions and data include all forms of non-volatile memory, including by way of example semiconductor memory devices, such as EPROM, EEPROM, and flash memory devices; magnetic disks such as internal hard disks and removable disks; magneto-optical disks; and DVD disks. Any of the foregoing may be supplemented by, or incorporated in, specially-designed application-specific integrated circuits (ASICs).

A number of embodiments of the invention defined by the following claims have been described. Nevertheless, it will be understood that various modifications to the described embodiments may be made without departing from the spirit and scope of the claimed invention. Accordingly, other embodiments are within the scope of the following claims.

Claims

1. A method for authenticating a message in a time division multiple access network, the method comprising:

receiving a message from a node at an active relaying component;

inserting a network-wide unique identifier into a field of the message at the active relaying component; and

forwarding the message on to other nodes in the network.

2. The method of claim 1, wherein inserting a network-wide unique identifier comprises inserting a unique identifier based on a port number of the active relaying component.

3. The method of claim 1, wherein receiving a message comprises receiving a message with a default value stored in the field of the message into which the active relaying device inserts the network-wide unique identifier.

4. The method of claim 3, wherein inserting a network-wide unique identifier comprises replacing the default value with a port number of the active relaying component at which the message was received.

5. The method of claim 1, and further comprising analyzing the message at the other nodes to authenticate the message.

6. The method of claim 5, and further comprising analyzing the message at the other nodes to detect passive behavior of the active relaying component.

7. The method of claim 6, wherein detecting passive behavior comprises detecting a short circuit in the active relaying component by absence of the network-wide unique identifier.

8. A method for authenticating a message in a time division multiple access network, the method comprising:

receiving a message from an active relaying component;

inspecting a value in the message inserted by the active relaying component; and

comparing the value with an expected value based on a transmission schedule.

9. The method of claim 8, wherein inspecting a value comprises reading a value derived from a port number at the active relaying device that received the message from a source node.

10. The method of claim 8, wherein inspecting a value comprises reading a value from a field in the message.

11. The method of claim 8, wherein inspecting a value comprises reading a value that is pre-pended or post-pended to the message.

12. A method for authenticating a message in a time division multiple access network, the method comprising:

receiving a message from a node at a active relaying component;

selectively performing a detectable modification to the message at the active relaying component based on local information; and

forwarding the message on to other nodes in the network.

13. The method of claim 12, wherein selectively performing a detectable modification to the message comprises one of inserting a value derived from a port number into the message, pre-pending or post-pending a value derived from a port number to the message and inverting the message.

14. The method of claim 12,

wherein selectively performing a detectable modification to the message comprises inverting the message; and

further comprising checking a port-derived value stored in the message by the node with a port number at which the message was received to authenticate the message.

15. The method of claim 12, and further comprising analyzing the message at the other nodes to authenticate the message.

16. The method of claim 15, and further comprising analyzing the message at the other nodes to detect passive behavior of the active relaying component.

17. The method of claim 16, wherein detecting passive behavior comprises detecting a short circuit in the active relaying component by absence of the detectable modification.

18. A method for authenticating a message in a time division multiple access network, the method comprising:

receiving a message from a node at an active relaying component;

appending a network-wide unique identifier to the message at the active relaying component; and

forwarding the message on to other nodes in the network.

19. The method of claim 18, wherein appending a network-wide unique identifier comprises one of appending the identifier to a leading edge of the message and appending the identifier to a trailing edge of the message.

20. A method for authenticating a message in a time division multiple access network, the method comprising:

receiving a message from a node at an active relaying component;

reading a value that uniquely identifies the port of the active relaying component that received the message;

verifying that the value corresponds to the port that received the message;

inverting the message at the active relaying component; and

forwarding the message on to other nodes in the network.

21. The method of claim 20, wherein reading a value that uniquely identifies the port of the active relaying component comprises reading a value from a field of the message.

22. The method of claim 20, wherein reading a value that uniquely identifies the port of the active relaying component comprises reading a value that is appended to the message.

23. The method of claim 20, and further comprising inverting the message at the other nodes to detect passive behavior of the active relaying component.

24. The method of claim 20, and further comprising inverting the message at the node prior to transmission to the active relaying component.

25. A machine-readable medium having instructions stored thereon for performing a method for authenticating a message in a time division multiple access network, the method comprising:

receiving a message from a node at an active relaying component;

selectively performing a detectable modification to the message at the active relaying component based on local information; and

forwarding the message on to other nodes in the network.

26. The machine-readable medium of claim 25, wherein selectively performing a detectable modification to the message comprises one of inserting a value derived from a port number into the message, pre-pending or post-pending a value derived from a port number to the message and inverting the message.

27. The machine-readable medium of claim 25,

wherein selectively performing a detectable modification to the message comprises inverting the message; and

further comprising checking a port-derived value stored in the message by the node with a port number at which the message was received to authenticate the message.

28. The machine-readable medium of claim 25, and further comprising analyzing the message at the other nodes to authenticate the message.

29. The machine-readable medium of claim 28, and further comprising analyzing the message at the other nodes to detect passive behavior of the active relaying component.

30. The machine-readable medium of claim 29, wherein detecting passive behavior comprises detecting a short circuit in the active relaying component by absence of the detectable modification.

31. A time division multiple access network, comprising:

a plurality of nodes;

an active relaying component;

each of the plurality of nodes coupled to the active relaying component over a communication medium; and

wherein the active relaying component performs a detectable modification to each message received from each of the plurality of nodes prior to forwarding the message to the other of the plurality of nodes.

32. The time division multiple access network of claim 31, wherein the active relaying component inserts a port-derived value in a field of each message.

33. The time division multiple access network of claim 31, wherein the active relaying component appends a port-derived value to each message.

34. The time division multiple access network of claim 31, wherein the active relaying component authenticates each message based on a port-derived value associated with the message and inverts each message prior to forwarding to the other of the plurality of nodes.

35. An apparatus for performing a method for authenticating a message in a time division multiple access network, the apparatus comprising:

means for receiving a message from a node;

means, responsive to the means for receiving, for selectively performing a detectable modification to the message based on local information; and

means, responsive to the means for selectively performing a detectable modification, for forwarding the message on to other nodes in the network.